Viruses and other malicious applications are increasingly being used to steal information from client devices, such as mobile phones, in an attempt to gain improper access to sensitive and protected client data. Conventional virus detection software often works by scanning downloads of files being stored on a client device to detect known viruses or malicious applications. In an alternative approach, malicious applications or viruses are detected behaviorally based on known common traits or signatures.
However, an issue with these known approaches is that the common traits, signatures, or definitions have to be known by the virus detection software. Further, the virus detection software has to include instructions for removing each specific malicious application. In other words, conventional known virus detection software is oftentimes reactive to because time is needed to analyze malicious applications prior to characterizing traits and determining a removal solution. Approaches to diagnose, remove, or contain such viruses and malicious applications without necessarily having identified the particular virus or virus signature are described in the present disclosure.